Table of Contents

• Setup
• Dataset
• Notes

Setup

• Set up a virtual environment

$ python -m venv env

• Install requirements

$ pip install -r requirements.txt 

Dataset

• This project uses the same dataset as the previous Mask R-CNN project in tensorflow. Your directory structure should be similar to the one shown below.

├───data
│   ├───test
│   ├───train
│   └───val
├───env
├───src
├───.gitignore
├────cardamage.ipynb
├───README.md
└───requirements.txt

• Run create_dataset in terminal

$ cd src
$ python create_dataset.py

Notes

• To install PyTorch on Windows, no CUDA:

  $ pip install torch==1.6.0+cpu torchvision==0.7.0+cpu -f https://download.pytorch.org/whl/torch_stable.html